Mapping artificial intelligence integration in higher education: a systematic review using the FACETS and SAMR frameworks

Abstract

Artificial intelligence (AI) is reshaping higher education through applications in teaching, learning, assessment, and curriculum design. Despite growing adoption, the literature remains fragmented, lacking structured frameworks to evaluate AI's integration and impact. This review aimed to map how AI has been integrated into higher education, classify applications using the SAMR model (Substitution, Augmentation, Modification, Redefinition), analyze reported outcomes and challenges, and apply the FACETS framework (Form, AI use case, Context, Education focus, Technology, SAMR) to enhance comparability across studies. The review followed PRISMA 2020 guidelines and was registered using the PRISMA-P protocol. Eight databases (PubMed, Medline, Web of Science, ProQuest, Scopus, Dimensions, OpenAlex, IEEE Xplore) were searched for English-language articles published between January 2015 and July 2025. Eligible studies examined AI integration in undergraduate higher education. A total of 959 records were screened in Rayyan QCRI. After removing duplicates and exclusions, 22 studies met the inclusion criteria. Data extraction covered study design, discipline, AI tool, SAMR level, outcomes, and FACETS dimensions. Most included studies originated from North America and Asia, with medicine, computer science, and engineering as leading disciplines. ChatGPT, was the most common platform. Applications clustered around assessment automation and personalized learning support. Most implementations were at the Substitution or Augmentation levels, with fewer Modification and one Redefinition. Reported benefits included efficiency, personalization, and engagement, while challenges were equity, ethics, and academic integrity. AI in higher education remains largely incremental, enhancing existing practices rather than transforming pedagogy. Its greatest potential lies in personalization.